Yards Per Target
Yards Per Target (YPT) has been added for WRs and TEs on the individual stat pages. It's exactly what it sounds like: receiving yards divided by how many passes were attempted to the receiver. It's hopefully something a little more useful than Yards Per Reception.
Posts
16 comments:
>>It's hopefully something a little more useful than Yards Per Reception.
Brian,
What is wrong with Yards per Reception, and how is this better?
Is it that this penalizes the deep receiver who when successful gains many yards, thus artificially inflating his YAR?
I think Air Yards, and Yards after Catch are rather useful. Any thoughts on integrating those into some sort of measurement?
- Jets Fan
It factors in catch rate, that's all. Nothing special. YPT includes some information about how well a receiver gets open and how well he makes a catch. YPR only considers successful plays, and doesn't consider bad routes, drops, etc.
Both stats, and all receiving stats, fail to isolate the receiver's sole individual component of success. QB and O-line obviously matter greatly.
How about yards*target? Yards are a good thing and targets are an indicator of quality. I know that in practice this list won't differ as much from a list of receiving yardage leaders, but I suspect that the stat (yards*targets) is more predictive of future success than yards/target.
Yards * target is no good because a high score will just show that someone is targeted a lot. This doesn't help us identify good receivers.
Interesting that most of the deep threats - rice, vincent jackson, floyd, etc are still on top of the list of yards per target. I guess in reality they should be since they are a higher risk play so they need a bigger reward.
But boy vincent jackson (or phillip rivers) sure was good this year. highest YPT, one of the highest deep %'s, and still a very high catch rate. That had to be one of the biggest snubs of the year for pro-bowl voting
Chase & 81Trucolors,
Actually, looks like a good list to me. Yes, it's people who are targeted a lot, but Chase's point is that there's a good reason they're being targeted; they're making the most of each one (hence the yards).
Here's the top 12: A.Johnson, R.Wayne, W.Welker, L.Fitzgerald, S.Rice, S.Smith (NYG), R.White, R.Moss, B.Marshall, S.Holmes, G.Jennings, M.Colston.
Admittedly, it looks a lot like the Yards leaderboard, with only Marshall and White pushing out V.Jackson and M.Austin. Though I'm not sure guys like Jackson should be penalized for not being targeted enough but still being a beast, as he is here.
While I like YPT, here's a stat I like even more: multiply YPT by %Tgt, and what you wind up with is Yards per Team Pass Attempt, which I think paints a better picture.
While it still gives credit for having the QB's trust (targets) and for production (yards), this "YPTPA" stat adds a twist by recognizing that it's harder to produce when you're the team's only offensive weapon, and thus draw more coverage. The YPT leaders generally play in great offenses that spread the ball around, making it that much easier to get open... I think taking into account the strength of the offense gives a better sense of a receiver's actual ability... what it gives you is simple, but effective: how much you're able to get open and produce every time your QB throws a pass.
The top 25: A.Johnson, S.Rice, V.Jackson, S.Holmes, S.Smith (NYG), A.Gates, H.Ward, S.Smith (CAR), J.Cotchery, W.Welker, R.Moss, G.Jennings, M.Austin, R.White, B.Marshall, D.Jackson, C.Ochocinco, D.Mason, R.Wayne, L.Fitzgerald, M.Colston, J.Witten, T.Owens, V.Davis, D.Driver.
OK, one more: I was curious how this would turn out with rushing, giving more credit to workhorse backs. In case anyone's interested, here are the '09 leaders in Yards per Team Rushing Attempt. Not too many surprises here:
C.Johnson, S.Jackson, M.Jones-Drew, F.Gore, A.Peterson, R.Rice, R.Grant, R.Mendenhall, J.Charles, F.Jackson, M.Forte.
Ryan and Chase-I like those ideas. The Yds per Team Tgt is the basis of an idea I've been kicking around. It could also be EPA per team attempt. There would be an adjustment for how many deep attempts each receiver had, since some WRs are asked to specialize in some route types.
There would be a baseline for each team to account (somewhat) for the strength of the QB and o-line, and then each receiver's score would be compared to his team's baseline. Another thought is to establish the team baseline on all pass attempts to receivers other than the receiver in question.
Brian,
One other thing I didn't account for, but maybe should, is QB sacks (i.e, use total "pass plays" instead of "team pass attempts") since a sack is no different from a dumpoff or throw-away in demonstrating the receiver's ability to get open. I thought about accounting for deep attempts but didn't know how; seems like Welker catching 3 of 4 short passes evens out with Moss's 3 of 5 with a lot more of them deep. How do you determine who's actually better?
I've also thought about comparing to other receivers on the same team as a baseline (also with RBs to see if running success is based on the line or the runner) but not sure how you'd normalize it; that is, Ward and Holmes comprise the top targeted duo in the league, but how much of that is because they're good, and how much is because the other Steeler WRs are crap, or they only run 2WR sets, etc? Do the Saints and Colts spread the ball around because they have the best QBs, or because all 4-5 of their receivers are great?
Awesome stuff. As always.
Question: what about RBs catching passes? You've probably considered it and maybe it's just hard to code in, but I was just wondering if that's coming down the line.
Thanks. Yup. It's easy to code. I'm just out of space in the table. I'm planning on at least two tables, one with conventional stats and another with 'advanced' stats. Working on defense stats now.
Great. Good to see that in there. So does WPA/EPA for RBs now include receptions?
Also, I noticed Jason Avant appeared as a RB on the Eagles 09 page.
Yes, WPA/EPA for RBs factors in their receiving performance. Likewise, WPA/EPA for WRs includes performance as runners, in reverse plays for example. It basically includes everything.
Regarding Avant, he was likely listed as a RB by mistake, or more likely an "HB" (H-back), in one of the official gamebooks. Right now, if I see HB, I have to call him a RB, since there are many teams that still refer to their primary RB as a halfback.
Soon I'll have a standard link where readers can report discrepancies like that. Thanks.
Avant sometimes lines up in the backfield when the Eagles get tricky near the goal line and put three guys behind Donovan. He typically motions out pre-snap, but that's probably what the scorer was seeing.
Yards-per-target is a great addition to the stat pages. I track and use several target-related stats when compiling my player projections. In fact, the first edition of my 2010 QB RB WR TE projections were released today. You can check them out here:
http://www.fantasydc.com/?page_id=167
Keep up the great work , Brian
Post a Comment